Evaluation of Face Recognition Technology

Facial recognition technology refers to a computer driven application that automatically identifies an individual from his or her digital image by a comparison of particular facial features in a facial database and in a live image (Vacca, 2007, 95).  The technology creates a template of peoples facial configurations, such as the lengths of their noses and the angles of their jaws.  It thereby functions like other biometric technologies (e.g. iris scanning) that use biological features for the purpose of recognition (Kautzer, 2002).  According to Visionics, a manufacturer of face recognition software, the technology is capable of finding human faces anywhere in the field of view and at any distance, and it can continuously track them and crop them out of the scene, matching the face against a watch list (Kautzer).  As an example of its application, facial recognition software is nowadays used for entrance management.  Systems are being designed to automatically open doors to allow only those individuals whose facial features are recognizable because they have been fed into those systems (Face Recognition Entrance Control System, 2009).  Undoubtedly, this application of face recognition software could secure both public and private buildings from criminals, including suicide bombers  the scare of the time.
But, is there more to face recognition technology than meets the eye  In other words, is it reasonable to rely on face recognition technology for entrance control apart from other applications related to security  Through this research paper, we would deepen our understanding of this technology so as to discover how the future of our criminal justice system can be positively impacted by it and our world made safer.  

Description and Uses of Facial Recognition Technology
It is natural brain technology that allows humans to recognize fellow human beings.  According to Bruce  Young (1986), there are face recognition units in the brains of human beings (361).  Face recognition involves matching the products of structural encoding, that is, facial features that are spotted by the eye with previously stored structural codes (Bruce  Young, 361).  Vacca writes that there are three areas of the face that are primary targets in face recognition because they do not typically change (95).  These parts of the face include upper sections of a persons eye sockets, the part of the face surrounding ones cheekbones, and the sides of the human mouth.  Regardless of whether the brain also considers these parts of the face most crucial to facial recognition, Vacca explains three different techniques used in man-made face recognition technology, namely, eigenface systems, eigenfeature systems, and thermal imaging.  Eigenface image systems capture facial images and change them to light and dark areas (Vacca, 95).  In eigenfeature systems, certain features of the face, for example, the eyes and the mouth, are picked out and distances are measured between these features.  Thermal imaging systems, on the other hand, take thermal images of the human face, focusing on the pattern of blood vessels (Vacca).
Seeing the problems with facial recognition in criminal trials, where eyewitnesses are given the responsibilities of facial recognition software, it easy to understand why complicated technological systems for face recognition are indispensable for the criminal justice system (Jong, Wagenaar, Wolters  Verstijnen, 2005, 87-88).  In fact, eyewitness identification depends on so many factors that even the United States Supreme Court has expressed doubts about its use in criminal trials, especially when eyewitnesses are shown photographs of the accused (Pezdek  Blandon-Gitlin, 2005, 247).  Biometric technologies, on the other hand, promise an accuracy rate of at least ninety-five percent (Hopkins, 1999).  Unsurprisingly, therefore, use of facial recognition software is widespread today (Bharvada, 2002, 270).
Even though iris scanning and other kinds of biometric technologies are known to be far more accurate than face recognition technology, the latter is more widely accepted because it is least intrusive.  This technology does not require users to push, click, or insert anything into the system, despite the fact that it takes many experts to create eigenface, eigenfeature or thermal imaging systems.  Moreover, organizations using the face recognition technology do not require the installation of anything except a new software application.  The cameras already in place as well as pictures of individuals on file are enough for organizations that use this technology.  Hence, face recognition technology is cheaper than iris scanning, for instance, which requires reading setups (Rutherford, 2001).  According to Frances Zelazney, who works as the director of corporate communication at Visionics, yet another advantage of facial recognition technology as compared to other biometric technologies is that facial recognition provides for inherent human backup because we naturally recognize one another  If the system goes down, someone can pull out an ID with a picture as backup, something you cant do with fingerprint devices (Rutherford).
Hence, facial recognition technology is known as the fastest growing biometric technology in our day.  Law enforcement agencies and the military have been successfully using the technology for many years without the public being aware of it.  As an example, in the year 1988 the Los Angeles County Sheriffs Department, Lakewood Division, began using composite sketches of suspects as well as video images in order to conduct searches on a database of digital facial shots.  The department also has a photo database of sex offenders and plans to find suspects on this database.  Then there is the Gang Reporting Evaluation Tracking system that can be searched with the use of photos of suspects in order for law enforcement to circumvent false identification cards as well as information that has been presented by gang members.  Other applications of face recognition technology in the area of criminal investigation include the Integrated Law Enforcement Face-Identification System which incorporates a unique three dimensional composite technology to identify angled-view face shots more easily.  The system is meant to be very helpful in the correct identification of uncooperative suspects in addition to subjects that have been caught from a distance using video surveillance cameras (Jarvis).
Britain is known to use at least two hundred thousand video cameras for surveillance.  Many of these cameras are being installed with facial recognition technology today.  Typically these systems use computers to monitor cameras that are looking for recognized criminals.  As soon as the system is seen to identify a known criminal, the police are called.  Numerous United States embassies around the world are also using face recognition systems to keep criminals from entering the country.  The Israel-Palestine border control is similarly using the technology to reduce crime (Jarvis).
Given the importance of putting a name to a face in law enforcement, whether it is to solve crimes, protect the public, or to ensure security in jails, face recognition software is proving itself to be of tremendous value.  Sheriff Everett Rice with the Pinellas County Sheriffs Office in Florida employs the technology to positively identify and verify individuals (Case Study Pinellas County Sheriffs Office Improves Law Enforcement).  Some of these individuals have been recently arrested while others are about to be released (Case Study Pinellas County Sheriffs Office Improves Law Enforcement).  So far, application of the technology has been successful, leading to hundreds of positive identifications of criminals (Case Study Pinellas County Sheriffs Office Improves Law Enforcement).  The Pinellas County Sheriffs Office has applied the new technology in other ways as well.  As an example, the Sheriffs Office has expanded use of face recognition systems to its patrol cars, allowing street deputies to identify those individuals that have been stopped and are unable to produce identification.  Although there is no estimate for how many criminals the law enforcement agency would catch through this system, users of the system are confident that face recognition software would have a greater impact on crime control in the years to come.  In fact, application of face recognition technology by the Pinellas County Sheriffs Office is a model for other agencies involved in criminal investigation (Case Study Pinellas County Sheriffs Office Improves Law Enforcement).  

Problems with Face Recognition Software
The United States Department of Defense, with its focus on perfecting face recognition technology to spot criminals on the borders of the nation, had been funding scientific studies on face recognition technology for more than a decade.  Private companies were similarly convinced that this technology could dramatically help in combating crime within the borders of the United States.  Because of their belief, marketing of facial recognition technology became widespread during the mid-1990s (Rutherford).
Then came 911  the day that changed the security concerns of the entire world in the matter of a few hours.  There was increased interest in face recognition systems following the terrorist attacks on American soil.  Although Americans had viewed face recognition technology with skepticism before the attacks, they became confident that pervasive use of the new technology in security as well as public safety would protect them from similar attacks in future.  Indeed, face recognition technology could play an important role in the prevention of tragedies.  All the same, law enforcement agencies have discovered that in areas covered by the new technology, no terrorist has ever been identified.  What is more, despite the redoubling of efforts to create dependable face recognition systems after 911, the new technology suffers from problems.  Facial recognition technology faces a difficulty, for example, in the recognition of the effects of aging.  Digitally compared photos of individuals that had been taken eighteen months apart produced untrue rejections by the software application at least forty three percent of the time.  Furthermore, it has been found that the technology is more successful when used by casinos to identify cheaters, in welfare offices, and by drivers license bureaus, given the uniformity of lighting and use of the same cameras in these places (Jarvis OHarrow, 2001).
Whether or not face recognition technology could drastically help in the areas of security and public safety is yet to be discovered.  Still, successful application of the new technology in the Pinellas County gives us an idea about the extent to which face recognition systems may enhance public safety management.  Perhaps, if all law enforcement agencies around the country were to employ face recognition technology, we would be able to confirm that the new technology may, indeed, work wonders for public safety.  Thus far, this is only an inference, seeing as common use of face recognition technology is not yet in vogue.
It may also be inferred that extensive use of the new technology around the world could save nations from the acts of terrorists.  Although face recognition technology is not as foolproof as public safety officials would like, it is an evolving technology incorporating new methods and systems of identifying the innocent as well as criminals.  In October 2001, the United States Senate subcommittee started to look into the potential uses of face recognition monitors in fighting terrorism (Rohde, 2001).  In point of fact, the government and the public at large have already understood that face recognition software may be very useful for security and public safety.  In the wake of the attacks on American soil, Viisage was requested to participate in a program at the Logan Airport to evaluate the potential of the new technology as a screening application for employees (QA with Viisage CEO Bernard Bailey).  The National Institute of Standards and Technology  playing a significant role in the enhancement of United States homeland security with projects that span a wide range of study areas and by helping the military, law enforcement, building and airport security besides other entities  has also developed a team of face recognition experts that work with federal agencies to improve computer applications with face recognition.  It has been reported that error rates in face recognition technology worked on by the National Institute of Standards and Technology dropped by almost fifty percent between the years 2000 and 2002 (Technologies for Public Safety and Security, 2007).
But, if technological experts are eventually able to address all problems related to face recognition software at present, it would still take training to perfect use of this technology.  Consider the case of Rob Milliron, a construction worker, who was enjoying his lunch in an entertainment area of Tampa, Florida, when a government camera equipped with face recognition technology took his photograph.  The photo was used without Millirons consent in an article published in the U.S. News  World Report.  When a woman in Oklahoma misidentified Milliron after seeing that photo and contacted the police department to have him arrested on child neglect charges, the man in the picture was forced to explain his innocence to law enforcement agencies (Alexander  Richert-Boe).  He told a newspaper once his explanation had been accepted They made me feel like a criminal (Alexander  Richert-Boe).

Although Millirons case raises ethical concerns regarding governmental use of facial recognition surveillance, it shows direction to both potential and existing users of this system in the effort to perfect its use.  After all, face recognition technology is essential to law and order in our day.  Richard Chace of the Security Industry Association stated upon seeing videotaped footage of a terrorist walking in the Logan Airport that if the security cameras had been equipped with the latest in face recognition technology, it would have been easy to prevent a hijacking (The Today Show, 2001).  However, Howard Simon, executive director of the Florida American Civil Liberties Union (ACLU), said that face recognition software might lead to racial profiling (ACLU Probes Police Use of Facial-Recognition Surveillance Cameras in Florida City, 2001).  However, there is no evidence that racial profiling has occurred due to the new technology.  Instead, there is widespread belief that face recognition systems would be of tremendous help in the security of the nation.
Indeed, face recognition technology is expected to go a long way in helping the nation and the world at large.  So therefore, improvements in the face recognition technology are awaited in order to put the new technology to better use than before.  Thus far, extensive use of the new technology has been impeded by scientific errors, for example, the fact that face recognition technology may not easily detect the effects of aging.  With new scientific research, however, such errors may be reduced and eventually done away with.  Pinellas County is already appreciative of face recognition technology.  Other law enforcement departments around the United States and around the globe may similarly find that it is better to put the technology to good use despite scientific errors than to avoid it altogether.  What is more, it must be remembered that there is no technology in the world that may be foolproof at all times.  Even law enforcement officers are error-prone.
It is for this reason that Garcia (2006) mentions the necessity of complementing facial recognition software with tamper detection sensors, line security supervision, door monitor switches, authorization verification checking, personal identification number checking, personal identity verification, personnel tracking, and two-door portals when this technology is used for entrance control (161).  As we have discovered through this discussion, face recognition software is not adequate on its own for entrance management because scientists have not been able to make it foolproof thus far.  Thus, organizations that purchase the Face Recognition Entrance Control System designed by FIRS Intelligent Technology (Shenzhen) Co. Ltd. or the likes of it must consider complementing the technology with other security check systems.  The Face Recognition Entrance Control System designed by FIRS Intelligent Technology sends a command for the door to open as soon as an individuals facial features are recognized (Face Recognition Entrance Control System).  But, this technology may or may not ensure security.  Then again, it depends on the security needs of the premises where face recognition software is installed for entry management.  Clearly, facial recognition technology would not ensure safety for all organizations that install it.  The fact remains, however, that it is better to use and perfect the use of this technology than to discard it altogether.


Post a Comment